Abstract
This paper investigates the impact of the widespread manipulation of reputation systems by sellers on two-sided online platforms. We focus on a relevant empirical setting: the use of fake product reviews on e-commerce platforms, which can affect consumer welfare via two channels. First, rating manipulation deceives consumers directly, shifting demand towards lower quality products through misinformation. Second, the presence of manipulation lowers trust that ratings reflect true quality, causing consumers to miss out on high quality products. Both misinformation and mistrust can also benefit consumers by increasing competitive pressure on products unable to differentiate through ratings. We study these effects by first modeling how consumers use ratings to form beliefs about quality and estimating consumers’ relevant priors about the prevalence of fake reviews using an incentivized survey experiment. We incorporate this beliefs model into a structural model of the Amazon marketplace that we estimate using a large and novel dataset centering on products observed buying fake reviews. We then use counterfactual policy simulations to evaluate how fake reviews impact the marketplace through misinformation and mistrust and explore the implications for platforms and regulators.